PRESSURE SEWER VS. GRAVITY SEWER: A CASE STUDY

1 1 PRESSURE SEWER VS. GRAVITY SEWER : A CASE STUDY Eamon Casey1 1. South East Water Corporation, Melbourne, VIC, Australia ABSTRACT This paper sets out the design journey incorporating significant innovations in the development of a backlog sewerage scheme for 16,200 properties in the Mornington Peninsula in Melbourne. From the initial considerations for a GRAVITY sewerage scheme the project evolved into the largest low PRESSURE SEWER network in Australia. The design utilised monitoring and control technology to optimise the network and overcome potential issues with power outage recovery and odour and corrosion management. INTRODUCTION The Peninsula Backlog SEWER Scheme in the Mornington Peninsula to the south east of Melbourne involves the connection of 16,200 properties to a reticulated sewerage scheme discharging to the existing Boneo Sewerage Treatment Plant (STP). The catchment is long and narrow with the most remote connection up to 17km from the Boneo STP.

2 The catchment is also characterised by undulating sand dune formations with associated difficulty in pipe installation. The initial high level feasibility assessment concluded that the environment prohibited a cost effective GRAVITY SEWER option and identified a low PRESSURE SEWER option being the optimal solution. This was on the basis that the GRAVITY SEWER option would require 27 transfer pumping stations and long lengths of GRAVITY SEWER with depths exceeding on narrow residential streets triggering micro tunnelling methodology. The high level cost estimate for the GRAVITY SEWER option exceeded $500M. The catchment is shown in Figure 1 with the individual reticulation catchment identified by colour and the transfer main alignment shown in blue. Figure 1 Mornington Peninsula Catchment Area Plan 2 METHODOLOGY/PROCESS PERCIEVED BARRIERS TO LOW PRESSURE SEWER The initial feasibility design took into consideration a number of barriers to low PRESSURE SEWER which dictated the scope of works for the overall system as follows: 1.

3 Self cleansing initial low flows 2. Pipe Sizing Peak Flow Analysis vs. Self cleansing 3. Power Outage Response 4. Odour & Corrosion 5. Power Consumption & Maintenance Costs The feasibility design for the low PRESSURE SEWER system addressed the above barriers and represented at a significant cost saving of $150M compared to the high level cost estimate for a GRAVITY option. Achieving this capital cost saving with a PRESSURE SEWER servicing strategy required an innovative design which optimised pipe sizes, the number & sizes of transfer pumps stations and the odour & corrosion control provisions. DESIGN EVOLUTION Two key innovations led to the evolution of the design to the current scope under construction. The first was the utilisation of dynamic hydraulic model which provides greater detail of system performance and critically an understanding of power outage response. The second innovation was the development of property pump monitoring and control through the ONEbox controller.

4 This allowed peak shifting, flushing cycles and reduced holding times to optimise asset sizing, assist odour control and management of power outage response. The low PRESSURE domestic pumps are provided with a high PRESSURE cut out facility based on high current monitoring. This has been set at 8 amps which is equivalent to pumping head of 45m to optimise the life of the low PRESSURE pumps. Since the low PRESSURE domestic pumps compete against each other for discharge to the PRESSURE reticulation network the PRESSURE head that each low PRESSURE pump sees is dependent on the number of low PRESSURE pumps in the local reticulation network that are operating. If a low PRESSURE pump cuts out on high current then the pump controller retries the pump after a set time period. The hydraulic model based on a maximum low PRESSURE pump head of 45m and an integrated reticulation and transfer main identified the range of the low PRESSURE pumps in the overall integrated network.

5 This low PRESSURE pump range identified by model became a key design factor in the optimum transfer pump station location. This design approach resulted in the number of transfer pump station being reduced to just 2. An additional key consideration in siting transfer pumping stations was to endeavour to maintain a fully pressurised network upstream of the break PRESSURE at the transfer pump stations in order to minimise air management requirements in the network. This was largely achieved by location of the transfer pump stations at local high points and utilisation of a barometric loop prior to discharge to the wet well to gain addition height. The increased head to pump over the barometric loop was considered a minimal issue when compared to the operational benefit associated with a fully pressurised network. The fully pressurised network provided consistent hydraulic performance and limits the critical management of odours to the transfer pump station sites.

6 3 There was an option to design transfer pumping stations as either in line booster pumps or a more traditional SEWER pumping station with discharge to a wet well and consequent break in PRESSURE . The decision was made to provide a wet well mostly on familiarity grounds for ease of operation. The overall network is therefore separated into 3 discrete integrated PRESSURE networks separated by a PRESSURE breaks at the 2 transfer pumping stations. The transfer pumps are designed for cut in/ cut out for low flows to achieve self cleansing velocities and follow the flow for higher flows utilising Variable Speed Drive (VSD). This approach reduces energy consumption and provides a more consistent flow to the treatment plant with consequent operational benefits. It is also noteworthy that the PRESSURE SEWER system is not restricted by incoming SEWER levels so the transfer pumping station can be located above ground if aesthetic consideration allow.

7 For one of the transfer pumping stations where visual impact can be managed the wet well is an above ground tank and the pumps are dry mounted. This has clear cost implications with reduced excavations but also operational benefits with dry mounting and ease of access to large pumps. With the scale of the integrated PRESSURE SEWER system and long retention times for the sewerage, the resulting anaerobic conditions lead to the reduction of sulphate in the sewage to sulphide. An odour model prepared for the system indicated dissolved sulphide levels of up to 26 mg S/L and low pH of The headspace Hydrogen Sulphide (H2S) levels at the transfer pump stations derived from the model was excessively high with average levels reaching 3500 ppm resulting directly from the high dissolved sulphide levels and low pH. Therefore the strategy for odour & corrosion control adopted was chemical dosing to control dissolved sulphide levels and air evacuation and treatment at the transfer pumping stations.

8 Dosing Ferric Chloride (FeCl3) upstream in the PRESSURE SEWER system effectively controlled dissolved sulphide levels below 1 mg S/L which resulted in reducing headspace H2S levels in the transfer pumping station to about 50 ppm. Dosing Magnesium Hydroxide (MHL) in addition to maintain pH above to ensure the effectiveness of the FeCl3 dosing was also adopted. Monitoring of pH during operation is intended to minimise the MHL dosing requirement. FeCl3 dosing was the most cost effective dosing option when compared to alternative liquid phase odour control measures. The air extraction and treatment system selected at the transfer pumping stations consists of a Bio Trickling Filter in series with an Activated Carbon Bed. The Bio Trickling Filter Treatment system was selected based on lower operating costs and robust performance based on the controlled range of H2S to be managed. The sizing of key assets including the transfer pumping stations and transfer mains was based on the model outputs for design flows and benefited from the complete removal of rainwater from the fully pressurised network and peak shifting using ONEbox based flow control.

9 The complete removal of rainwater can be achieved with pump monitoring and comparison with rainfall data for infiltration on the customer side of the low PRESSURE pump. Figure 2 shows actual data from South East Water s network identifying rain infiltration indicated by higher frequency of pump runs during a rainfall event which can be subsequently targeted. Figure 2 RIn additicapacitydomestiback flowperiods pressureachievab TARGETThe finamodellinSelf cleaPeak FloFigure 3 Rainfall Infiltration the diurny for sizing ofic pump podw during peaor peak shife SEWER catcble reductionTED SOLUTIOl design addng and ONEbansing initialThe ONEboxcatchment amultiple prooptimising tintermittentconnectionsow Analysis &Peak Shifting Dtion Identificatnal flow proff assets. Withs facilitates hak periods anfting is utilishment over n in peak flowONS TO PERC ressed each box control al low flows x controller pare controlleoperty pumphe system fot flushing cyc.

10 & Asset SizinData ion Utilising Floile associateh a controlleholding packnd draw dowsed to manag24 hours witw is by a factCIEVED LOW of the low pas follows: provides a flud through ans pumping toor ultimate pcles for self cng 4 ow Monitoringed with seweed low pressuk flow duringwn levels in tge flows. Figth and withotor of 2. PRESSURE Spressure sewushing moden algorithm together. Thispeak demandcleansing at ig rage flows gure system tg peak flow pthe domesticure 3 below out peak shifSEWER BARRwer barrier isse whereby alto maintain as facilitated td with largerinitial stagesenerally dicthe storage aperiods. Contc pump podsshows actuating control RIERS sues with thel connectiona desired peathe design tocatchment s with a limitetates the minavailable in thtrolling flowss during low fal data from in operatione benefit of ns within a suak flow witho focus on thsizes and proed number o nimum he s to hold flow a n. The ub he oviding of 5 The peak flows in the final design are taken from the dynamic model which are based on monitored dry weather data from elsewhere within SE Water s network and therefore represent a robust design.